Leaked DeepSeek V4-Pro Materials Reveal Coding Performance Near Claude at 1/50th the Price
ccerrato147 · x · 2026-08-07
Leaked investor materials suggest that DeepSeek V4-Pro will enter the global top tier. Its coding capabilities reportedly lag Anthropic's flagship model by only 0.3%, while supporting a 1M token context window.
Based on Anthropic's flagship benchmarks, V4-Pro is estimated to score extremely close on Terminal-Bench 2.1, DeepSWE, and SWE-bench Pro. The API pricing gap is even more staggering: for a mix of 1M input and 1M output tokens, V4-Pro costs roughly 1/46th of Anthropic's flagship. Its cache hit price is allegedly 1/276th of the competition, making the claim of being '10x to 100x cheaper' highly plausible.
Related event: DeepSeek Resumes $8B Funding, V4-Pro Specs Leak(8 posts)→
More from Models
- Ornith-1.5 Open Models Released, Claiming Claude Opus Performance — alejandroll10 · 2026-08-26
- 14-year AI veteran: Grok understood code I thought no one ever would — Kuprel · 2026-08-26
- Together Ranks Top Open Models: Kimi K3 and DeepSeek V4 Lead Use Cases — togethercompute · 2026-08-26
- Questions over Astra's progress: 2 months for 3 more models? — teortaxesTex · 2026-08-26
- View: Tokens-per-second matters more than model size now — natesiggard · 2026-08-26
- Tiel-Coder-35B achieves 121.4 tok/s for local inference — DerTomsn · 2026-08-26